Bob Friday Talks: The AI-Native Networking Platform Defined

0:00 hello AI native networking platform has

0:03 been something we've talked about a lot

0:04 lately but what does that mean for your

0:06 user and your operator to help us

0:08 understand it all we have brought in Bob

0:10 Friday our chief AI officer to walk us

0:12 through it all right Bob first question

0:15 what role does AI play in today's

0:17 networks and how is AI able to benefit

0:19 the experience of the end user yeah

0:22 Jessica you know for me I want to relate

0:24 a story from when I started Miss I was

0:26 talking to some very large retail

0:28 customers and what they were telling me

0:30 was Bob before I put any business

0:32 critical app onto this network you have

0:34 to assure me that your controllers are

0:36 not going to crash that you can

0:37 basically release code more than once or

0:39 twice a year and most importantly you

0:42 have to make sure that when that user

0:44 connects that they're going to have a

0:45 great experience right and that's when I

0:48 realized that really the Paradigm was

0:50 shifting from helping customers manage

0:52 Network elements to really helping them

0:54 manage that cloud AI user experience and

0:58 that is the role that AI is playing in

1:00 today's network is really moving from

1:02 that Paradigm of managing Network

1:03 elements to managing the client to Cloud

1:06 user experience all right so what is the

1:08 difference between AI native and AI

1:11 driven that is a great question you know

1:14 my story was at Cisco before I started

1:16 Mist I was talking to some very critical

1:19 big retail customers and they told me a

1:21 couple things one is before they were

1:24 going to put any business critical app

1:25 on that Network like an app for their

1:27 customers they wanted to make sure that

1:29 controllers were not going to crash

1:31 anymore they wanted to make sure that I

1:33 could deliver code more than once or

1:36 twice a year they wanted me to keep up

1:37 with their mobile app and more

1:39 importantly they really wanted to make

1:41 sure before anyone on that Network that

1:44 they were going to have a great

1:45 experience and that was really the

1:48 paradigm shift from managing Network

1:50 elements like I did 20 years ago versus

1:53 managing the unit experience so when you

1:55 look at AI native the decision I made

1:58 was to solve that problem

2:00 I need to leave Cisco and start with a

2:02 clean sheet of paper you know that

2:05 actually building an AI native was going

2:07 to require a real-time Foundation Cloud

2:10 solution that you were not going to get

2:12 at a legacy system and more importantly

2:15 when you put the software on the cloud

2:17 you need to build a microservices

2:19 software architecture which really

2:21 required that blank sheet of paper so

2:23 when you think about AI native and AI

2:25 driven AI native means a clean sheet of

2:28 paper and a Foundation built for

2:31 real-time Cloud AI performance Aid

2:34 driven is basically make taking that

2:36 foundation and then delivering an aid

2:38 driven experience so with AI native I've

2:41 been hearing part of the reason Juniper

2:43 AI is ahead is because it's built on the

2:46 right data the right real-time response

2:49 and the right secure infrastructure so

2:51 can you explain as to what that means

2:53 you know because when I was at Mist the

2:56 reason I built an access point was to

2:58 make sure I could get the right data

3:00 to answer the question of why is a user

3:02 having a poor internet experience it's

3:05 not because I thought the industry

3:06 needed another access point it's because

3:08 I wanted to make sure I could get the

3:09 right data in terms of real-time

3:11 response the other reason I left Cisco

3:14 to start Miss because I knew that we had

3:16 to build a cloud infrastructure a secure

3:18 Cloud infrastructure that really could

3:20 handle data in real time and when you

3:22 look at what the first thing we built

3:24 indor location we were able to take data

3:27 back from the network device and

3:30 actually return a location response in

3:32 one second so that's what we mean by

3:34 realtime data down to second level

3:36 real-time data across a secure Cloud

3:39 infrastructure so Juniper offers the

3:41 only platform with a common

3:43 microservices Cloud why is that

3:45 important for AI and how does that help

3:47 deliver an exceptional user experience

3:50 you know as I mentioned before when I

3:51 was working with large retail customers

3:53 one of the things they asked for was for

3:55 controllers to stop crashing and that is

3:58 really due to the difference the

3:59 software architecture between an on-

4:01 Prim solution and a Cloud solution you

4:04 don't can't simply take software on your

4:06 controller and move it to a container or

4:08 Cloud that does not funly change the

4:10 reliability when we move to these Cloud

4:12 architectures we're moving to a

4:14 microservices architector that we can

4:16 basically download and release every

4:19 other week which brings the speed of

4:20 innovation and on terms of reliability

4:24 these microsurfaces Cloud architectures

4:26 reduce the blast radius over in the

4:28 controller on Prim world we had

4:30 spaghetti coat so we're fundamentally

4:32 changing the software architecture from

4:34 a very spaghetti Cod type environment to

4:37 a much more structured Cloud API

4:39 environment where microservices reduce

4:41 the blast radius and allows the speed of

4:44 innovation all right so Juniper just

4:46 announced two enhancements to Marvis why

4:50 should every Enterprise be leveraging

4:51 Marvis minis technology we're very

4:54 excited about Marvis minis you know up

4:56 to now we've done a great job on the

4:58 pre- connection experience and the post

5:00 connection user experience what Marvis

5:02 minis really allows us to do is to start

5:04 to bring data to Marvis when there are

5:07 no users on your network this will allow

5:09 our customers to make sure that that

5:11 network is ready to go on that first day

5:13 of deployment or when that network is

5:15 ready to go when they open the doors at

5:16 8:00 so marvous minis think of it as a

5:19 synthetic user that's constantly on your

5:22 network constantly making sure your

5:23 network is ready for prime time all

5:26 right so we discussed two enhancements

5:28 talking about the second one Juniper

5:30 extends Marvis to the data center why is

5:33 this integration a game-changing

5:35 Innovation and what real world benefits

5:37 should it leaders expect since joining

5:40 Juniper in 2019 we've been on the

5:42 mission here to extend Cloud Ai and

5:45 Marvis across the Enterprise portfolio

5:47 wireless switch route now we're looking

5:50 to extend Marvis across the different

5:53 business units here and campus branch

5:55 and data center is one of those

5:57 Integrations and why is this important

5:58 to our customer customers because we

6:00 have several customers that run business

6:02 critical apps like point of sale

6:04 boarding pass out of their private Data

6:06 Center and when those critical business

6:08 apps go down they need to quickly figure

6:11 out is it a campus land branch problem

6:14 or is it really just a data center

6:15 problem so by combining both the branch

6:18 and the data center into Marvis now

6:20 we're helping our customer answer

6:21 business critical application

6:23 experiences all right so last tough

6:26 question how does the Personnel change

6:29 when is introduced yeah so you look at

6:31 it departments and how they've evolved

6:33 over the last 20 years we're kind of

6:35 moving from a era of CLI the dashboards

6:40 and now we're moving on to Virtual

6:42 assistance llm and chat GPT so when you

6:45 look at the IT personnel as we start to

6:48 introduce cloud and AI into the mix

6:51 first of all we're starting to see it

6:52 move from clii to becoming python

6:55 programmers right apis are becoming much

6:57 important we're seeing that right now as

6:59 we speak as we move forward we're going

7:02 to start to see llm conversational

7:04 interfaces become much more important

7:06 those it persons and we're going to see

7:08 them start to interact with their

7:09 networks in a much more natural way they

7:11 asked us three years ago to find this

7:13 problem we actually just found it now

7:15 with a new AI model that basically is

7:18 able to predict Zoom user experience and

7:20 what it showed was they basically had a

7:22 misconfigured VPN Gateway that was

7:24 routing traffic from India to Australia

7:28 right and so this is the power of AI is

7:30 helping it Enterprises find those needle

7:34 in the Hast stack problems so this is an

7:36 exciting time to be back into networking

7:38 again and Cloud AI is going to make it

7:39 funner for the next

7:46 decade